Meeting the lead-free inspection challenge
Steve Scheiber, Contributing Technical Editor -- Test & Measurement World, 8/1/2005
The movement away from using lead-based solders presents challenges to board manufacturers. Not only must they find a solder that offers properties as close as possible to those of lead solder, but they must also find new ways to test their products and verify their circuits. Automated inspection offers one answer to the second problem.
Regulations in Europe and China require the removal of all leaded components and solder in certain electronic products sold there by July 1, 2006. Other countries—including the US—are expected to follow suit.
The new regulations do not cover all products. Exceptions include telecommunications network infrastructure, medical products, and defense and aerospace systems. Nevertheless, high-volume products such as personal computers, televisions, DVD players, and cellular phones must conform. Stig Oresjo, senior test strategy consultant with Agilent Technologies, estimates that the regulations will cover 60–70% of all boards sold into the regulating countries (Ref. 1).
Although companies might like to make the transition by following Path A in Figure 1, limited knowledge of the effect of lead-free solder on the process and the dearth of lead-free components will make that approach impractical. Industries forced to go to lead-free will probably follow Path B, incorporating lead-free solder and adding lead-free components as they become available.
 |
| Fig. 1 While most manufacturers would prefer to make the transition to lead-free manufacturing by following Path A, they are more likely to follow Path B. Manufacturers not governed by the regulations may follow Path C. Courtesy of Agilent Technologies. |
Path C will apply more to exempt industries, where companies will continue to use conventional tin/lead solder but will adopt lead-free components as parts makers introduce them. Companies that make both covered and exempt products may elect to make all their products lead-free to avoid maintaining two different manufacturing processes, inventories, and test and inspection strategies. New alloys, new failures
After experimenting with metals, the electronics industry has settled on a range of alloys consisting mainly of tin with a few percent silver and less than 1% copper. These solutions prove less than perfect, however. Board manufacturers can expect more defects, at least at first. Lead-free solder melts at 217°C compared to 183°C for its predecessor. The higher temperature required for soldering and rework will subject components and substrates to additional stresses that will reduce their reliability.
Oresjo contends that because of the higher temperature, manufacturers may permit fewer repair attempts after a failure. "Shotgunning" a board after a functional-test failure may prove damaging as well. Both procedural changes will likely mean more scrap.
Also of concern, thanks to its metallurgy rather than its melting temperature, is that the new alloy exhibits lower wetting characteristics compared to lead solder. Wetting of molten solder determines how well it covers pads and device leads, as well as the shape of the resulting solder joints. Sufficient wetting also helps slightly misplaced components to adjust themselves by floating a little on pads during reflow.
Wetting forces during reflow tend to draw tin/lead solder paste back to pads, reducing the likelihood of shorts caused by solder bridges. Lead-free boards won't react in the same way, and will therefore tend to have more shorts. Also, lead-free wave-soldering may produce insufficient barrel fill on boards with through holes.
Because some manufacturers have started to conform to the regulations, the industry already has data on lead-free manufacturing. Figure 2 shows results from one manufacturer who makes two similar boards—one using tin/lead solder and the other lead-free. Each board contains about 1300 components and about 3000 solder joints, including BGAs, gull-wings, and SMT connectors, with the smallest component an 0402 with a pitch as low as 20 mils. The data cover 85,000 leaded boards and 60,000 lead-free. The figure includes only fault types affected by the change.
 |
| Fig. 2 One contract manufacturer reports these defect levels for high-volume PCBs; the results cover 85,000 leaded boards and 60,000 lead-free boards. Courtesy of Agilent Technologies. |
The greatest increase appears in tombstone faults, with significant increases in solder opens and chip misalignments. Contrary to expectation, these data show a decline in defect rates for solder bridges and shorts. Oresjo suggests that this result may be mere coincidence or may result from the close attention paid to the manufacture of this board by experienced engineers who kept the process under tighter-than-normal control. Impact on inspection
From a process perspective, making lead-free boards will require more accurate solder-paste deposition to avoid inadequate pad coverage. Although inadequate pad coverage would not generally be flagged as a defect, a fixture nail might miss the solder during in-circuit test and hit bare copper. Because solder provides better nail contact than copper does, the resulting tests may generate false failures. Similarly, the inability of components to correct themselves during reflow will mean stricter calibration of pick-and-place machines.
To deal with these issues, Oresjo recommends performing automated optical inspection (AOI) at post-paste and post-placement. Post-paste AOI will ensure adequate solder-pad coverage and allow for inexpensive correction. Correcting inadequate or inconsistent solder volume before adding components will dramatically reduce solder-related problems later on.
Post-placement AOI would find misaligned components. With leaded solder, this step is often skipped precisely because the devices may move during reflow. Since movement is less likely with lead-free solder, a post-placement defect will remain after reflow, and repairing it at the earlier step proves less expensive.
Because of its lower wetting properties, lead-free solder is also less forgiving of slightly bent device leads, which may contribute either to actual performance faults or reliability (field) failures. Again, an additional inspection step can identify these problems even if the circuit functions properly.
Unlike other test methods, AOI can handle lead-free alloys with little adjustment. The new alloy is less shiny than the leaded version, joint shape will differ somewhat, and lead-free joints are grainier than with traditional solder, but such differences are relatively minor. Oresjo cites a study conducted by the National Physical Laboratory in the UK on six AOI systems. It concluded that the AOI step will perform as well as before with few modifications.
The effect of the higher defect levels that lead-free solder will likely create, however, may prove more problematic. Manufacturers already contend that AOI generates too many false calls, especially false failures. With a rise in board failure rates, such complaints will likely become more insistent, even if the inspection results are as reliable as they were before.
Production bottlenecks may present a problem. Even cursory AOI takes time, and an increase in inspection efforts to cope with rising defect levels may mean inspection times that exceed the permissible beat rate. In addition, since a manufacturer may perform AOI at more places in the process, this problem may occur at more points as well.
Reliability concerns will increase the need to perform post-reflow inspection. Lead-free solder is more brittle than tin/lead, so even if a circuit functions in the factory, there is no guarantee that shipping it to the customer will not cause a failure. To handle BGAs and other hidden nodes, automated x-ray inspection (AXI) will likely become the post-reflow inspection method of choice.
Lead-free solders offer less impedance to x-rays. Recalibration of AXI equipment can compensate. If both leaded and lead-free boards of the same type come down the line, for example, merely tweaking gray levels in the test program may suffice.
Analysis of inspection results offers side benefits. Diligently feeding results back into the process can permit improvements that minimize future defects. When manufacturers first turned to no-clean processes several years ago, defect levels in many places rose by an order of magnitude. As the industry gained experience in how the process differed from its predecessor, defect levels began to drop until they reached close to the level where they had stood before the change. Oresjo suggests that lead-free manufacturing will likely follow a similar pattern.
Making the transitionBefore addressing the issues raised by lead-free manufacturing, Oresjo recommends looking carefully at your current defect levels and fault spectrum. Understanding where you are today will help you manage the transition. He also suggests switching over only one manufacturing line at a time so you can more easily track and solve problems. Each line switchover should then proceed more smoothly than the last, and the entire migration will prove as painless as possible.
| Reference |
|
